Simultaneous Equation Error

[husnafezaahmad]

Hello everyone,

I have a problem to run my model. Initially it was OK, but once i put a shadow variable as an input to this one variable, and poof! the error came out.
I had read the solution in previous same topic as mine here, but i don't think my problem is the same.

Attached is the screenshot of my model, as well as the .mdl file of it.

Your help is much appreciated.

[husnafezaahmad]

The .mdl file.

I left it in my previous post.

Hello everyone,

I have a problem to run my model. Initially it was OK, but once i put a shadow variable as an input to this one variable, and poof! the error came out.
I had read the solution in previous same topic as mine here, but i don't think my problem is the same.

Attached is the screenshot of my model, as well as the .mdl file of it.

Your help is much appreciated.

[Administrator]

but once i put a shadow variable as an input to this one variable, and poof! the error came out.

Which two variables are you referring to?

[Administrator]

I'm guessing that you are using the wrong equation for "new coal". You are using the STEP function, I suspect that you want a delay here instead.

[husnafezaahmad]

OMG! Yes, i want a delay over there. Thought that STEP can also be treated as delay thanks! I’ll work on it and i really pray it’ll work.

I'm guessing that you are using the wrong equation for "new coal". You are using the STEP function, I suspect that you want a delay here instead.

[tomfid]

I think the .mdl attachment was omitted.

Looking at the diagram, it seems like there's an issue with the regression. I'm not sure whether the "load growth" stock has units of energy (additional load above some baseline) or energy/time (rate of change of load). But both seem problematic:

additional load {kw} = f( population {people}, GDP {$/yr} )

might be ok, but could be non-robust if (for example) low values of the inputs cause negative load.

load growth rate {kw/year} = f( population {people}, GDP {$/yr} )

is really problematic because it's mixing stocks and flows, i.e. regressing through an integration.

[husnafezaahmad]

Here is my file. Dunno why it was omitted.

Since i'm so new in this vensim world, i'm kinda confused tho with the suitability for one variable to either be stock or not.

I think the .mdl attachment was omitted.

Looking at the diagram, it seems like there's an issue with the regression. I'm not sure whether the "load growth" stock has units of energy (additional load above some baseline) or energy/time (rate of change of load). But both seem problematic:

additional load {kw} = f( population {people}, GDP {$/yr} )

might be ok, but could be non-robust if (for example) low values of the inputs cause negative load.

load growth rate {kw/year} = f( population {people}, GDP {$/yr} )

is really problematic because it's mixing stocks and flows, i.e. regressing through an integration.

[husnafezaahmad]

I’m still struggling with this simultaneous error.

Can anyone help me?

[tomfid]

The version attached above runs for me - not clear how to replicate the error.

The issue is likely a feedback loop without a stock in it. I think this model would benefit from a complete rework of the stock-flow usage. Start by going through and getting rid of the boxes on anything that is not a stock (level) and doesn't use a function with an internal level (SMOOTH or DELAY functions).

The regression definitely seems problematic. Consider what happens if population and GDP are constant: load grows at a constant rate. Why would that happen?